|Home | About | Journals | Submit | Contact Us | Français|
Chemokine receptor mediated recruitment of inflammatory cells is essential for innate immune defense against microbial infection. Recruitment of Ly6Chi inflammatory monocytes from bone marrow to sites of microbial infection is dependent on CCR2, a chemokine receptor that responds to MCP-1 and MCP-3. While CCR2-/- mice are markedly more susceptible to L. monocytogenes infection than wild type mice, MCP-1-/- mice have an intermediate phenotype, suggesting that other CCR2 ligands contribute to antimicrobial defense. Herein, we show that L. monocytogenes infection rapidly induces MCP-3 in tissue culture macrophages and in serum, spleen, liver and kidney following in vivo infection. Only cytosol invasive L. monocytogenes induce MCP-3, suggesting that cytosolic innate immune detection mechanisms trigger chemokine production. MCP-3-/- mice clear bacteria less effectively from the spleen than WT mice, a defect that correlates with diminished inflammatory monocyte recruitment. MCP-3-/- mice have significantly fewer Ly6Chi monocytes in the spleen and bloodstream, and increased monocyte numbers in bone marrow. MCP-3-/- mice, like MCP-1-/- mice, have fewer TNF- and iNOS-producing dendritic cells (Tip-DCs) in the spleen following L. monocytogenes infection. Our data demonstrate that MCP-3 and MCP-1 provide parallel contributions to CCR2-mediated inflammatory monocyte recruitment and that both chemokines are required for optimal innate immune defense against L. monocytogenes infection.
Chemokines play a central role in homeostatic trafficking and positioning of immune cells and also provide guidance and direction to immune cells responding to inflammatory challenges (1). By binding to chemokine receptors and triggering intracellular signaling pathways that induce cell movement (2), chemokines orchestrate many aspects of innate and adaptive immune responses. For example, CCR2 and its chemokine ligand MCP-1, play pivotal roles in monocyte recruitment during infection or under other inflammatory conditions(3-11) and CX3CR1 and its ligand CX3CL1 (Fractalkine) facilitate monocyte and macrophage trafficking in atherogenesis (12, 13). CCR2, CCR5, CXCR3 and CX3CR1 contribute to rapid NK cell mobilization that occurs in inflammatory conditions (14-18). CCR7 and its ligands, CCL19 and CCL21, direct the trafficking of T cells, B cells (19, 20) and activated dendritic cells (DCs) (21) to lymph nodes (LN), a process that promotes priming and activation of naïve T cells (22, 23) and colocalization of B and T cells within LNs (24, 25). CCR5 and its ligands facilitate efficient CD8 T cell priming by promoting migration of naïve CD8 T cells towards mature, antigen-bearing dendritic cells(26).
Cellular trafficking is a complex process. Directionality is provided by chemokine gradients that are established within tissues by interactions between chemokines and glycosaminoglycans (GAGs) (27, 28). Although GAG association is not required for in vitro chemokine activity, mutations within human MCP-1 that impair GAG binding inhibit MCP-1 mediated monocyte recruitment in vivo (29). Because chemokines differ in their affinity and specificity for tissue GAGs, it has been postulated that GAGs, by selectively binding distinct chemokines, play an essential role in directing leukocyte trafficking(30). Association with GAGs can also promote chemokine oligomerization, which also promotes chemokine-mediated leukocyte recruitment (28, 31). In addition to GAG association and oligomerization, shear forces have also emerged as a factor contributing to chemokine activity (32). In a low-flow, shear-free environment, surface-bound CCL21 promoted sustained T cell motility without activating integrins. However, when shear forces were applied, integrins were activated by CCL21 and resulted in arrest of T cells. Thus, this mechanism allows chemokines to trigger cell extravasation at specific sites in the vasculature, under high flow conditions, while enhancing the mobility of trafficking cells in static compartments.
Chemokine and chemokine receptors play important roles during bacterial infection. L. monocytogenes is a Gram positive intracellular bacterial pathogen that causes severe disease in immunocompromised hosts(33). L. monocytogenes virulence depends upon the pore-forming protein listeriolysin O (LLO), which enables bacterial escape from the phagosome and entry into the host cell cytoplasm(34). Early immune defense against L. monocytogenes infection requires CCR-2 mediated recruitment of inflammatory monocytes from the bone marrow (11) and differentiation of recruited monocytes into TNF- and inducible nitric oxide synthase (iNOS)-producing dendritic cells (Tip-DCs) at sites of infection (35). In mice lacking CCR2 or MCP-1, Tip-DC recruitment is markedly diminished in spleen, resulting in enhanced in vivo bacterial growth that overwhelms the host (10).
Although the importance of CCR2 in immune defense against L. monocytogenes infection is well established, less is known about the relative contributions of CCR2 ligands to innate immune defense. In addition to MCP-1, three other CC-chemokines, MCP-2, MCP-3 and MCP-5, can bind CCR2 and trigger signaling. MCP-3 was identified as a protein with monocyte chemotactic activity that copurified with MCP-1 from stimulated human osteosarcoma cells (36). MCP-3 is structurally and functionally similar to MCP-1 and is a potent in vitro chemoattractant for monocytes(36), T cells (37) and NK cells (38). Recently, an MCP-3-/- mouse strain was generated and demonstrated that MCP-3 serves as an important CCR2 ligand that enhances in vivo monocyte recruitment (39).
In the present study, we characterize the expression of MCP-3 during L. monocytogenes infection, and investigate the relative contributions of MCP-3 and MCP-1 to monocyte recruitment and innate immune defense. We demonstrate that MCP-3 is rapidly induced following L. monocytogenes infection and that MCP-3-/- mice, like MCP-1-/- mice, are more susceptible to in vivo infection. Increased susceptibility of MCP-3-/- mice correlates with decreased inflammatory monocyte recruitment from bone marrow into circulation. Our data demonstrate, for the first time, that MCP-3 plays a critical role in immune defense against L. monocytogenes infection by mediating inflammatory monocyte recruitment.
All mice used in this study were bred at Memorial Sloan-Kettering Research Animal Resources Center. Generation of MCP-1-/-, MCP-3-/- and CCR2-/- mice was previously described (4, 39, 40). All mice were backcrossed at least 10 generations onto the C57BL/6 background. Mice were infected intravenously with 3000 L. monocytogenes strain 10403S, 5× 105 attenuated ActA-deficient L. monocytogenes strain DP-L1942, 108 attenuated LLO-deficient L. monocytogenes strain DP-L2161, or 109 HKLM prepared as decribed previously(41). At indicated times following infection, spleens were harvested and dissociated in phosphate-buffered saline containing 0.05% Triton X-100 and bacterial CFUs were determined by plating on brain-heart infusion agar plates.
Macrophages were grown from bone marrow precursors in antibiotic-free DMEM medium supplemented with 30% supernatant from L-cell fibroblasts and 20% fetal bovine serum. On day 5, cells were harvested in ice-cold PBS, plated at 2-4 × 105/well in 96-well flat bottom plates, and various bacterial strains or 108/ml HKLM were added to the cells. Bacteria were grown to log phase (A600 of 0.1), pelleted at 10,000 rpm for 10 min, and resuspended in PBS prior to addition to cells. Wild type and LLO-deficient bacteria were added to the cells at 5:1 bacteria:cell ratio. Total volume in the wells during infection was 200 μl in 96-well plates. Infection was allowed to proceed for 30 min at which time extracellular bacteria were washed away, and gentamicin-containing media was added to each well to prevent extracellular bacterial growth. Supernatants were collected 2, 4 and 6hr postinfection to assay for chemokines.
At various times after infection, spleens were collected and dissociated and were digested with 0.3% collagenase type 4 (Worthington). Bone marrow cells were collected from mouse femurs. The following antibodies were purchased from BD Pharmingen (San Diego, CA): anti-CD11b-PerCP(M1/70), anti-Ly6C-FITC(AL-21), anti-TNF-FITC(MP6-XT22). Goat anti-iNOS antibody (M-19) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and FITC-anti-goat IgG was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). For FACS analysis, a large FSC/SSC gate was drawn to include lymphocyte/monocyte populations. Intracellular staining for iNOS was performed by staining cells for cell surface markers, fixing in 2% paraformaldehyde, permeabilizing with Perm/Wash Buffer (BD Pharmingen), and incubating with goat anti-iNOS antibody followed by FITC-anti-goat IgG. For intracellular TNF staining, splenocytes were stimulated in vitro with 108/ml HKLM in the presence of brefeldin A (BFA) for 4hr and processed according to manufacturer’s protocol (Cytofix/Cytoperm, Pharmingen).
Murine MCP-1 was quantified using an ELISA kit from BD (San Jose, CA), and murine MCP-3 were quantified using an ELISA kit from Bender MedSystems (Burlingame, CA). To obtain lysates for chemokine assays, organs were harvested at indicated times following infection, macerated in ice-cold PBS containing 0.01% Triton X-100, and centrifuged at 10,000 × g.
Our previous studies demonstrated that MCP-1 expression contributes to CCR2-mediated monocyte recruitment (10) during L. monocytogenes infection and mice lacking CCR2 or MCP-1 are highly susceptible to L. monocytogenes infection (3, 10). While the expression of MCP-1 following L. monocytogenes infection has been characterized, very little is known about the expression of MCP-3. Therefore, as a first step, we investigated the in vivo induction of MCP-1 and MCP-3 following L. monocytogenes infection. As shown in Figure 1, we found that the kinetics of MCP-1 and MCP-3 induction are similar, with dramatic induction of both chemokines in spleen, serum and liver 24hr and 48hr following infection. We noticed that there was higher level of MCP-1 than MCP-3 in the spleen 48hr postinfection. Induction of MCP-3 in spleen, serum, liver and kidney was maximal 48 hours following infection. Although the kinetics of MCP-1 and MCP-3 induction share many similarities, some significant differences were also detected. For one, while MCP-1 levels decrease between 48hr to 72hrs of infection, MCP-3 levels in spleen and liver remain elevated. In addition, levels of MCP-3 in the kidney are quite high while levels of MCP-1 in the kidney remain relatively low throughout infection. Whether this disparity reflects differences in expression or trapping of these chemokines in the kidney requires further investigation.
Bacterial invasion of host cell cytosol is required for MCP-1 induction by L. monocytogenes (10). To determine whether MCP-3 expression is also induced by an intracellular sensing mechanism, we used an attenuated L. monocytogenes strain that lacks the LLO gene for in vitro and in vivo infections. As shown in Figure 2A, MCP-3 production was markedly reduced in vivo when the mice were innoculated with LLO-deficient L. monocytogenes or heat-killed Listeria (HKLM), while infection with ActAko bacteria, which invade the cytosol but are attenuated because they are unable to spread from cell to cell, induced MCP-3. For in vitro analyses, BMMØs were infected with wild-type or LLO-deficient L. monocytogenes, or stimulated with HKLM. Minimal MCP-3 induction was detected when macrophages were infected with LLO-deficient bacteria or exposed to HKLM (Figure 2B). These experiments, therefore, suggest that MCP-3 and MCP-1 are both induced by cytosol invasion by live L. monocytogenes.
CCR2 signaling is pivotal for innate immune defense against L. monocytogenes infection. While several monocyte chemoattractant proteins (MCPs) can bind to CCR2, MCP-1 and MCP-3 were shown to be major ligands for CCR2 in inflammatory settings (39). To examine the role of MCP-3 in innate immune defense against bacterial infection, we compared susceptibilities of wild-type C57BL/6 mice and mice lacking CCR2, MCP-1 or MCP-3 to L. monocytogenes infection. Figure 3A demonstrates that MCP-1-, MCP-3- and CCR2-deficient mice had increased numbers of L. monocytogenes in spleens 3 days following bacterial inoculation when compared to the wild-type mice. The increased susceptibility of MCP-1-, MCP-3- or CCR2-deficient mice was even more apparent 5 days following infection (Figure 3B). Between day 3 and day 5 of infection in wild-type mice the number of viable bacteria decreased by 80%, while the number of bacteria in MCP-1-/- and MCP-3-/- mice remained similar during this time period. In contrast, the number of bacteria in spleens of CCR2-/- mice increased 500 fold between day 3 and day 5 of infection. The similar intermediate phenotype of MCP-1-/- and MCP-3-/- mice indicates that MCP-1 and MCP-3 both contribute to innate immune defense against L. monocytogenes infection.
The frequency of circulating inflammatory monocytes is diminished in CCR2-deficient mice. The extent to which MCP-1 and MCP-3 contribute to inflammatory monocyte emigration from bone marrow in the absence of infection is unclear. Therefore, we determined the frequency of inflammatory monocytes in spleen, blood and bone marrow of naïve wild-type, MCP-1-/-, MCP-3-/- and CCR2-/- mice. We identified inflammatory monocytes by their high level expression of Ly6C, and intermediate to high level expression of CD11b. In Figure 4, flow cytometric analysis revealed that the percentage of CD11bint/Ly6Chigh cells was significantly reduced in spleens (Figure 4A) and blood (Figure 4B), and retained in the bone marrow (Figure 4C) of naïve MCP-1-/-, MCP-3-/- and CCR2-/- mice. Compared to MCP-1-/- mice, MCP-3-/- mice had a greater reduction of circulating monocytes, suggesting that MCP-3 may, under non-inflammatory circumstances, play a more important role in maintaining monocyte homeostasis.
Previous studies characterized a novel dendritic cell subset, Tip-DCs, which are recruited to the spleen in a CCR2-dependent manner and which serve as a major source of TNF and iNOS during early L. monocytogenes infection (35). To determine whether the increased susceptibility of MCP-3-/- mice correlates with a defect in Tip-DC recruitment following infection, we examined Tip-DC populations in spleens of MCP-3-/- mice. The level of CD11bint/Ly6Chigh Tip-DCs is markedly decreased in the spleens (Figure 5A) and blood (Figure 5B) of infected MCP-1-/-, MCP-3-/- and CCR2-/- mice. The magnitude of the decrease in Tip-DC recruitment correlates with the increased susceptibility of mice to L. monocytogenes infection, with intermediate numbers of TipDCs in MCP-1-/- and MCP-3-/- mice compared to wild-type and CCR2-/- mice. These results suggest that MCP-1 and MCP-3 each play a partial role in monocyte trafficking during infection. Thus, MCP-1-independent monocyte recruitment is likely mediated by MCP-3 signaling through CCR2, while MCP-3-independent recruitment is likely mediated by MCP-1. However, an MCP-1-/-MCP-3-/- mouse strain is not currently available and thus it is not possible to determine whether MCP-1 and MCP-3 account for all CCR2-dependent inflammatory monocyte recruitment.
Reduction in Tip-DC recruitment to the spleen and blood may result from the failure of monocytes to emigrate from the bone marrow or from abnormal trafficking of monocytes to other tissues. Recent studies demonstrated that CCR2-mediated signals are required for inflammatory monocyte emigration from bone marrow into the circulation (11). To determine whether MCP-3 is required for monocyte emigration from the bone marrow in the setting of L. monocytogenes infection, we examined the CD11bint/Ly6Chigh monocyte population in bone marrow of wild-type, MCP-1-/-, MCP-3-/- and CCR2-/- mice 24hrs following infection. We detected an increase in CD11bint/Ly6Chigh cells in the bone marrow of MCP-1-/-, MCP-3-/- and CCR2-/- mice (Figure 5C). Similar to the observation we had in naïve mice, MCP-3-/- mice had a more severe defect in monocyte emigration from bone marrow during early L. monocytogenes infection than MCP-1-/- mice. Defects in monocyte recruitment in MCP-1-/- and MCP-3-/- mice were most dramatic during early infection. Differences in monocyte frequencies between wildtype and MCP-1-/- or MCP-3-/- mice, though significant, were diminished 48hrs after infection in spleen (Figure 5D), blood (Figure 5E) and bone marrow (Figure 5F).
To understand whether MCP-1 and MCP-3 have roles in activation in addition to recruitment of inflammatory monocytes, we examined TNF-α and iNOS expression by inflammatory monocyte population in the spleen. Intracellular cytokine staining results showed that TNF-α (Figure 6A) and iNOS (Figure 6B) expression levels in CD11bintLy6Chi cells in MCP-1-/- and MCP-3-/- spleens were similar to wild-type mice, indicating MCP-1 and MCP-1 are not required for Tip-DC activation. Consistent with results we obtained by CD11b and Ly6C surface staining, we also observed diminished recruitment of CD11bintTNF-αhi and CD11bintiNOS+ cells in MCP-1-/-, MCP-3-/- and CCR2-/- mice. These results suggest that higher susceptibilities of MCP-1-/- and MCP-3-/- mice are due to decreased recruitment of inflammatory monocytes rather than impaired differentiation of recruited monocytes into Tip-DCs.
Recruitment and targeting of inflammatory cells during infection is a complex process that is enabled, in part, by chemokine mediated signals. The aggregate inflammatory response depends on recruitment of many different cell types that either circulate in the bloodstream or reside in depots throughout the body, such as spleen, lymph nodes and bone marrow. Among the cytokines and receptors that are available to coordinate this process are the roughly 50 distinct chemokines and 20 different chemokine receptors encoded in the mammalian genome. Some chemokine-chemokine receptor pairs are monogamous, such as CXCL12-CXCR4 and CCL20-CCR6. Other chemokines, however, bind to multiple receptors, such as RANTES(CCL5) with CCR1, CCR3 and CCR5. On the other hand, a number of chemokine receptors bind multiple chemokine ligands, as demonstrated by CCR7’s association with CCL19 and CCL21, and CCR2’s response to MCP-1, MCP-2, MCP-3 and MCP-5. Although CCR2 responds to multiple ligands in vitro, the relative contribution of each ligand to cellular trafficking and immune defense during bacterial infection has remained largely unexplored.
In this study, we show similar induction kinetics and expression levels of two CCR2 ligands, MCP-1 and MCP-3, following L. monocytogenes infection. Expression of either MCP-1 or MCP-3 depends on bacterial invasion of the host cell cytosol. Both MCP-1 and MCP-3, under uninflamed conditions, promote homeostatic emigration of inflammatory monocytes from bone marrow into circulating blood. We demonstrate that MCP-1 and MCP-3 both contribute to emigration of monocytes from the bone marrow and Tip-DC recruitment into the spleen during L. monocytogenes infection. Lack of either chemokine results in impaired clearance of bacteria from the spleen and enhanced susceptibility to infection.
Inflammatory chemokines are rapidly induced following infection, and bind to their receptors to initiate cellular migration and activation. Recruiting inflammatory cells expeditiously and directing them sites of infection so they can kill invading microbes is a core responsibility of the innate immune system. Precise regulation of chemokine expression, receptor association and subsequent cellular responses likely optimizes protection and minimizes deleterious effects of uncontrolled microbial growth and invasion. It remains unclear, however, how expression of multiple chemokines that bind the same receptor contributes to innate immune defense. A simple hypothesis is that expression of several ligands for a critical receptor provides redundancy in the event one ligand is either lost or inactivated. In this scenario, the function of these different ligands would be essentially identical. An alternative but more complex model is that different chemokines binding the same receptor are not simply redundant but provide distinct signals, perhaps due to variations in their association with the chemokine receptor, their ability to dimerize/oligomerize or their association with GAGs. In this setting, different chemokines would provide distinct information to responding cells, either at the signaling level or at a spatial level, depending on where the chemokine binds. Some evidence in support of the second model comes from studies of CCR7-mediated cell activation and recruitment. While CCL19 and CCL21 both bind CCR7, they provide distinct signals to responding cells, as detected by in vitro disparities in receptor desensitization, receptor phosphorylation, and ERK1/2 activation (42). In vivo, CCL21 is sufficient for the proper organization of lymphoid T cell zones and migration of lymphocytes, including transmigration across HEVs, while CCL19, but not CCL21, contributes to homeostatic maintenance of T cells (43).
In our study, we observed similar defects in monocyte trafficking and similarly increased CFUs in MCP-1-/- and MCP-3-/- mice following L. monocytogenes infection. In comparison to CCR2-/- mice, MCP-1-/- and MCP-3-/- mice have intermediate phenotypes and thus, MCP-1 and MCP-3 appear to contribute nearly equivalently to monocyte recruitment and antimicrobial clearance. It is possible that MCP-1 and MCP-3 make purely redundant contributions to monocyte recruitment. In support of this scenario, our studies find similar kinetics of MCP-1 and MCP-3 induction and roughly similar amounts of chemokine in most tissues. On the other hand, our studies also detected subtle differences in the duration of MCP-1 and MCP-3 expression in tissues and a marked increase of MCP-3, but not MCP-1, protein levels in kidneys of infected mice. Whether these differences in chemokine levels have functional implications will require further study. In addition, it is still possible other CCR2 ligands such as MCP-2 and MCP-5 play a role in the monocyte recruitment process following infection, although they are not implicated in homeostatic monocyte emigration (39).
Characterization of human MCP-1 and MCP-3 demonstrated differences between these two chemokines that might impact their activities in vivo. Human MCP-3, for example, binds to heparan or heparan sulfate as a monomer, while human MCP-1 only associates as a dimer (44, 45). If MCP-1 and MCP-3 make distinct contributions to monocyte recruitment, it is possible that these two chemokines function in parallel, each controlling the movement of a subset of monocytes, or in series, each controlling the movement of the whole monocyte population at distinct stages of their migration from bone marrow to the site of infection. Although our results indicate that MCP-1 and MCP-3 both function at the level of monocyte emigration from the bone marrow, since the frequencies of monocytes in the bone marrow of MCP-1-/- and MCP-3-/- mice were similar, neither the “in parallel” or “in series” model can be excluded. Thus, it is possible that two populations of monocytes reside in the bone marrow, one responding to MCP-1 and the other to MCP-3 (Fig 7A). Alternatively, it is possible that MCP-1 directs the movement of monocytes from one location in the bone marrow to another, and that MCP-3 provides an additional stimulus that enhances movement into the circulation (Fig 7B). Distinguishing between these models will likely require intravital studies of monocyte trafficking within bone marrow.
In summary, MCP-1 and MCP-3 have additive roles in monocyte recruitment and host defense against L. monocytogenes infection. Low secretion of MCP-1 and MCP-3 under normal conditions maintains homeostatic migration of inflammatory monocytes from bone marrow into the circulation. During infection, this process is enhanced by high levels of MCP-1 and MCP-3 in the spleen and serum, or by local production of these two chemokines in the bone marrow. It is still not clear where MCP-1 and MCP-3 are produced and where they establish chemokine gradients in vivo following infection. Further investigation of the expression and activity of these two chemokines during infection will be valuable to dissect the respective roles of these two CCR2 ligands.
The author’s research is supported by National Institutes of Health (R37AI03903, E.G.P; R01HL52773 and R01HL63894, I.F.C.).